Do As I Say: Curricular Discordance in Medical School End-of-Life Care Education

JOURNAL OF PALLIATIVE MEDICINEVolume 10, Number 3, 2007© Mary Ann Liebert, Inc.DOI: 10.1089/jpm.2006.0190

Do As I Say: Curricular Discordance in Medical School End-of-Life Care Education

MICHAEL RABOW, M.D.,1 JOHN GARGANI, Ph.D.(c),2 and MOLLY COOKE, M.D.3

To be good is noble; to teach others to be good is nobler—and less trouble.

—Mark Twain

ABSTRACT

Purpose: Prior research indicates that medical students face significant personal and ethical chal-lenges when they perceive differences between what is taught in class (formal curriculum) and whatis practiced on the wards (informal curriculum). This study seeks to further describe the educa-tional experience and characteristics of students who perceive curricular discordance in end-of-lifecare (EOLC) training.

Method: Self-administered questionnaire among third-year medical students at a large, urbanmedical school. Statistics to identify the correlates of perceived curricular discordance.

Results: Completed surveys were returned by 141 students out of a class of 162 (response rate �86.5%). Student perception of curricular discordance was inversely correlated with student per-ception of educational quality (p � 0.001) and their regard for institutional values (p � 0.001). Stu-dent attitudes and emotional reactions did not correlate with curricular discordance, nor did stu-dent age, gender, ethnicity, or prior personal experience with death of a loved one. Students hadlimited role modeling and feedback. While student informal curricular experiences did not corre-late with a perception of curricular discordance, an increased number of informal curricular EOLCexperiences did correlate with a perception of poorer educational quality (p � 0.05).

Conclusion: Curricular discordance is correlated with a perception of poorer educational qualityand more negative regard for institutional EOLC values, but not correlated with other features ofstudent experience or characteristics. Importantly, increased informal EOLC experiences are asso-ciated with a more negative assessment of educational quality.

759

INTRODUCTION

WHAT STUDENTS ARE TAUGHT in class is not nec-essarily what they learn. Medical students likely

learn the knowledge, skills, and attitudes of medicalprofessionals via a combination of formal teaching and

informal influences, including the combination of un-official messages, interpersonal relationships, and ed-ucational social milieu often referred to as the “infor-mal” or “hidden” curriculum.1–4 Although students areexposed to informal curricular influences from thevery beginning of medical school, clinical rotations in

1Department of Medicine, University of California, San Francisco, San Francisco, California.2Gargani � Company, Berkeley, California.3Department of Medicine, University of California, San Francisco, San Francisco, California.

RABOW ET AL.760

which students observe supervisors in practice are re-plete with these informal educational experiences.During third-year clerkships, students confront the hid-den curriculum as they enter into new relationshipswith classmates, house staff, and faculty, witness peersand working professionals model emotional reactions,navigate the social environment and organizationalstructure of their school, and decipher institutional ex-pectations, values and priorities that prescribe theirconduct.1–3,5–8

Informal and formal curricula may be concordantor discordant with each other. Although informal cur-ricular messages may be aligned with formal curric-ula, given the realities of classroom versus clinical ex-perience, students typically face discrepanciesbetween their formal and informal curricular teach-ings. Although the discrepancies could go in either di-rection, most formal curricula present a principled andidealized approach to a topic while the actual imple-mentation of care on the wards is often fraught withcompromise, conflicting priorities, and practical limi-tations. Although the impact of curricular discordanceon the learning of content knowledge is unknown,prior research has suggested that seeing supervisorspractice contrarily to the professional values and ethicsofficially recommended in school has deleterious con-sequences for students’ commitment to ethical princi-ples and their emotional well-being.4,9–16

However, there have been only a few empiric stud-ies of the informal curriculum and how its interactionwith explicit, formal teaching affects students andlearning. Education in end-of-life care (EOLC) affordsan excellent opportunity in which to study the hiddencurriculum and curricular discordance17 becauseEOLC training typically involves discrete formal pre-

sentations,18–22 as well as intense experiences, emo-tions and relationships.20,23 Addressing the hidden cur-riculum in EOLC (in addition to simply developingmore EOLC content) has been advocated as a meansof improving EOLC education.24–27

The purpose of this study is to describe empiricallyinformal EOLC training experiences at one medicalschool and to assess possible correlates of perceivedcurricular discordance with data from a survey of third-year medical students. Although the determinants oflearning are likely to be numerous and highly com-plex, to begin to explore the correlates of curriculardiscordance, we propose a simple conceptual model(Fig. 1) in which formal and informal curricular ex-periences interact to create a perception of curriculardiscordance, which then influences the medical stu-dent learning experience. Although formal and infor-mal curricular influences likely impact each other, ourstudy looks simply at the correlates of perceiving dis-cordances between them. The learning experience in-cludes such elements as students’ assessment of thequality of their EOLC training, their regard for theEOLC values of their teaching institution, their atti-tudes about EOLC, and their emotional experience indealing with EOLC. Prior research has suggested thatage, gender, ethnicity, and prior experiences with thedeath of a loved one also may be relevant to EOLClearning.28–31 Such student characteristics (demo-graphic variables, prior personal experiences) may in-fluence whether students perceive discordance in cur-ricular massages.

We hypothesize that discordance between the for-mal and informal curriculum leads students to a morenegative assessment of the quality of their EOLC ed-ucation, a more negative regard for their school’s

FIG. 1. Conceptual model of correlates of perceived curricular discordance. (EOLC, end-of-life care.)

CURRICULAR DISCORDANCE IN END-OF-LIFE CARE EDUCATION 761

EOLC values, and more negative EOLC attitudes andemotions. Noting prior literature suggesting a negativeimpact of curricular inconsistencies on student well-being, we expect that the impact on students likely in-volves their emotional experience of EOLC. With un-comfortable emotions and threats to well-being,student may be predisposed to develop negative atti-tudes about EOLC as well. Although no prior workhas examined the impact of curricular discordance onstudents’ estimation of their school and faculty, giventhat the school and faculty are the source of the dis-cordance, we predict that the negative impact of dis-cordance extends beyond students’ experiences ofthemselves to include students’ perception of theschool as well. Additionally, we hypothesize that stu-dent characteristics and prior EOLC experiences im-pact the likelihood of their perceiving curricular dis-cordance, with older, female, and white students, andstudents who have experienced grief in the past beingmore likely to perceive discordance.

METHODS

This is a descriptive and correlational study. Giventhat students have had their most intense experienceswith both the formal and informal curriculum by the endof their first year of inpatient ward rotations, third-yearmedical students attending University of California, SanFrancisco (UCSF), were chosen to complete a writtensurvey to examine the correlates and impact of curricu-lar discordance. By the end of second year, all respon-dents had completed 15 hours of required EOLC didac-tic, small group, and preceptorship experiences. Thesewere followed by traditional inpatient clerkships duringthe third year. Students completed the survey during thetenth month of their third year. Confidentiality was main-tained via randomly generated numerical coding of eachsubmitted questionnaire. Respondents received the in-centive of a chance to win 1 of 20 copies of a best-sell-ing medical textbook for completion of the question-naire. The study was approved by UCSF’s InstitutionalReview Board prior to data collection.

The survey was self-administered and includedquestions about classroom and clinical EOLC trainingexperiences. The survey included Likert scale and di-chotomous response options, and its development wasinformed by a literature review, prior EOLC and edu-cation surveys,20 and review by national experts in hid-den curriculum theory and end-of-life care education.The survey was pilot-tested with second-year studentsat an affiliated medical program and then revised toimprove its structure and clarity.

The survey explored multiple elements of studentidentity and experience, including student demographiccharacteristics, informal curricular experiences (includ-ing patient interaction, role modeling, and feedback),perception of discordance between the formal and in-formal curriculum, EOLC attitudes, EOLC emotions,perception of educational quality, and regard for institu-tional EOLC values. Specific survey items addressingeach of these elements are shown in Appendix A. Theimpact of the formal curriculum was excluded from theanalysis because all respondents participated in the sameformal curricular offerings.

An independent research firm double-entered sur-vey responses by hand. These data were then analyzedusing Minitab and Conquest statistical software. Com-parisons between groups were analyzed with t tests.Pearson correlations were calculated to assess the as-sociation between scaled variables and student per-ception of curricular discordance. Significance wastaken at p � 0.05.

For the analysis of correlations between the variousscales, each student was assigned a Rasch measure for(1) regard for institutional EOLC values, (2) EOLCemotions, (3) EOLC attitudes, (4) discordance be-tween the formal and informal curriculum, and (5) in-formal EOLC curriculum. Similar items were bundledtogether as composite scales indicating the level ofeach investigated construct possessed by respondents.For each of these scales, a scaled score was estimatedfor every respondent using an extension of the Raschmodel, which is commonly employed by researchersin education and psychology.32 In this study, the Raschmodel is particularly useful for its ability to combineordinal survey responses to produce interval levelscores suitable for regression analysis.33 Correlationsbetween constructs were estimated as correlations be-tween scaled scores (without adjustment for attenua-tion). Survey responses were also analyzed at the itemlevel, in which case they are reported as averages onthe Likert response scales for ease of interpretation.Because there was only one question associated withperception of educational quality, a Rasch measurecould not be assigned. The range of these Rasch mea-sures is given in Table 1, while Likert scores are re-ported in the text below.

RESULTS

Response rate and student characteristics

Of 162 possible third-year respondents, 141 stu-dents completed surveys (response rate � 86.5%).The average age of respondents was 26.4 years old.

TABLE 1. VARIABLES, SCALES, AND SCALE RELIABILITY

Number of Scale Scale ReliabilityVariable or scale Meaning survey items range mean SD �

Student characteristics

Age Student age in years 1 23 to 41 26.42 2.65 NAEthnicity Student race/ethnicity 1 NA NA NA NAGender Student gender 1 NA NA NA NAPersonal EOLC Whether student had 1 NA NA NA NA

Experience experienced the death ofa loved one

Curricular influences: Informal EOLC curriculum

Patient interaction Student observation of or 3 0 to 10a 4.1413 2.0520 0.620interaction with dyingpatients. Higher valuesindicaate more frequentpatient interactions.

Role modeling Student observation of or 2interaction with attendingphysicians or housestaffaround EOLC. Highervalues indicate morefrequent interactionswith supervisors.

Feedback Clinical feedback from 1attending physician ofstudent EOLC. Highervalues indicate morefrequent feedback fromsupervisors.

Percepton of curricular discordance

Perception of curricular Student perception of 4 0 to 15a 8.6736 2.5647 0.623discordance differences between the

formal and informalcurricula. Higher valuesindicate a sense of greaterdiscordance.

EOLC learning experience

EOLC attitudes Student attitudes about 4 5 to 10a 0.0343 0.9760 0.206the importance of EOLCto them as learners andclinicians. Higher valuesindicate feeling morestrongly about doctorcompetence and studentexperience with EOLC.

EOLC emotions Student emotional 5 5 to 15a 10.2868 1.714 0.484response to observing orproviding EOLC. Highervalues indicate morenegative emotions aroundEOLC.

Perception of overall Student report on the quality 1 1 to 5 2.3357 0.7258 NAeducational quality of their overlal EOLC (where 3 is

education. Higher values “average”)indicate a perception of higher quality.

Regard for institutional Student report on the value 3 0 to 20a 11.8838 3.5956 0.652values UCSF exhibits for EOLC

and dying patients. Highervalues indicate a higherperceived value of EOLC at UCSF.

aIndicates score that was rescaled using a Rasch measure.EOLC, end-of-life care.

More than half (57%) were female. Half of studentsidentified themselves as white (50%), and one quar-ter identified themselves as Asian (25%). The re-mainder identified themselves as African American,Hispanic, other or multiple ethnicities. Most (76%)students reported personally having suffered the lossof a loved one.

The informal curriculum in EOLC

Informal EOLC curricular experiences commonlyincluded interactions with patients, but less commonlyincluded role modeling and feedback from clinical su-pervisors. Specifically, almost all third year students(134 students, 95.7%) reported interactions with pa-tients at the end of life. However, role modeling waslimited: 51 (36.2%) never observed an attending physi-cian break bad news and 89 (63.1%) never saw an at-tending conduct an advance directive discussion. Stu-dents reported that their own clinical discussions withattendings about EOLC were limited, with formal orinformal discussions occurring between “rarely” and“sometimes” during the care of dying patients (mean �2.57 [standard deviation {SD} � 1.4] on a Likert scalewhere 1 � never, 2 � rarely, 3 � sometimes, 4 � usu-ally, and 5 � always). Students were less likely to haveEOLC discussions with attending physicians than withhouse staff (mean Likert scores � 2.6 and 2.85, re-spectively; p � 0.001). When discussions did occur, at-tending physicians were less likely to talk about theirown emotions around EOLC than were house staff (p �0.008). One third of students (32.7%) reported neverhaving received feedback from supervisors about anyof their EOLC clinical activities.

The EOLC learning experience

Overall, students assessed the overall quality of theirEOLC education as below “average” (mean � 2.32,SD � 0.73 on a reversed Likert scale where 1 � ter-rible, 3 � average, and 5 � outstanding). Studentsgenerally agreed that their medical center valuedEOLC (mean � 3.97, SD � 0.071 on a 5-point Likertscale where 1 � strongly disagree and 5 � stronglyagree with the statement that “End-of-life care ishighly valued at UCSF”). For EOLC attitudes, studentswere on average “neutral” (mean � 3.30, SD � 1.39on a Likert scale where 1 � strongly disagree, 3 �neutral, and 5 � strongly agree). With respect toEOLC emotions, students on average were “neutral”about feeling badly or guilty about their interactionwith EOLC patients (mean � 2.80, SD � 1.20 on aLikert scale where 1 � strongly disagree, 3 � neutral,and 5 � strongly agree).

CURRICULAR DISCORDANCE IN END-OF-LIFE CARE EDUCATION 763

The correlates of perception of curricular discordance

Survey scale descriptors and measure reliability aregiven in Table 1. Correlations between perception ofcurricular discordance with student characteristics, in-formal curriculum, and EOLC learning experience areshown in Table 2.

There were no significant correlations between stu-dent perception of curricular discordance and the stu-dent characteristics analyzed, including age, gender,ethnicity, or personal experience with the death of aloved one. There were no significant correlations be-tween student perception of curricular discordance andinformal curricular experiences. However, informalcurricular experiences were inversely correlated withstudent perception of educational quality (p � 0.05).The more EOLC experiences a student had, the morenegatively they rated the quality of their EOLC train-ing. Students who had the greatest number of infor-mal EOLC experiences, including interaction with pa-tients and role modeling and feedback from clinicalsupervisors, tended to rate the quality of their EOLCeducation more negatively.

Student perception of curricular discordance wassignificantly negatively correlated with both studentperception of educational quality (p � 0.001) and stu-dent regard for institutional EOLC values (p � 0.001).Curricular discordance was not significantly correlatedwith EOLC attitudes or EOLC emotions.

DISCUSSION

Among medical students toward the end of theirthird year, the greater the discrepancy they perceived

TABLE 2. CORRELATIONS OF PERCEPTION

OF CURRICULAR DISCORDANCE

EstimatedPearson

Subscale correlations p value

Age 0.12 0.24Malea �0.17 0.08Personal EOLC experiencea �0.013 0.90Ethnicitya

White �0.003 0.97Asian �0.046 0.64Hispanic 0.143 0.15Other �0.052 0.559

Perception of educational quality �0.284 0.001Regard for institutional values �0.318 �0.001EOLC attitude 0.067 0.431EOLC emotions 0.139 0.10

aDichotomous indicator variable.EOLC, end of life care.

between what was taught in the classroom and howthey saw EOLC practiced on the wards, the more neg-atively they rated the quality of their EOLC educationand the value their school put on EOLC. However,perceiving curricular discordances was not associatedwith more negative EOLC attitudes or emotions. Stu-dent characteristics such as age, gender, ethnicity andprior experience with the death of a loved one alsowere not correlated with perception of curricular dis-cordance. While student exposure to the informalEOLC curriculum, including patient experiences, clin-ical role modeling and feedback, did not correlate withperception of curricular discordance, greater exposureto informal EOLC curriculum was negatively corre-lated with student perception of educational quality.

Our empiric findings add to what is known aboutthe determinants and consequence of perceiving cur-ricular discordance. When students see greater incon-sistencies between what they have been taught in classand how they see their teachers actually practice onthe wards, they offer a lower assessment of the qual-ity of their training and a greater sense that their in-stitution does not genuinely care about EOLC. Al-though prior work suggests that curricular discordanceand observing professional lapses lead to an erosionof ethical values and threaten student well-being,4,9–16

we did not find evidence that curricular discordance iscorrelated with student EOLC attitudes or emotions.In our study, the negative consequences of curriculardiscordance appear to be externalized onto the insti-tution, rather than internalized by the student. Little isknown about the impact on students’ actual behavioror their professional practice once they have completedtheir training.

Also contrary to what some prior literature might sug-gest, our data indicate that basic student characteristicsare not correlated with student perception of curriculardiscordance. Although older students are sometimesthought to bring greater maturity and perspective to theirwork,28 age was not correlated with perception of dis-cordance here. Although gender has been suggested asan influence on whether students may be “immunized”against adopting tacit, non-virtuous values,29 gender didnot correlate with recognition of curricular conflicts inthis empiric study. Although student ethnicity has beenshown to bear on student attitudes relevant to EOLC,30

our analysis indicates that ethnicity did not correlatewith perception of curricular discordance. Finally, al-though one might expect personal EOLC experiencesto sensitize students,31 students who had suffered thedeath of a loved one were no more likely to perceiveconflicts with how EOLC was discussed in class andpracticed on the wards.

Notably, in our study, more patient interaction, con-tact with faculty, and feedback were associated with amore negative assessment of the quality of EOLCtraining. Our data do not offer a definitive explanationfor this finding, but it is possible that student supervi-sion and role modeling were so poor in quality that theimpact of informal curricular experiences on educa-tion was negative. Our study is consistent with othersthat have demonstrated that students have limited ex-posure to attending physician emotional role model-ing.7,11,20,23,34 However, the quality of patient contactand role modeling experiences and a student’s abilityto process them, rather than just their frequency, likelydetermine whether students experience these interac-tions as valuable. Given the complexity and intensityof EOLC, students may require sophisticated supervi-sion in processing, understanding, and learning fromtheir EOLC experiences.35 These results offer a noteof caution about the tendency to respond reflexivelyto curricular EOLC deficiencies by simply adding di-dactic content and clinical experiences.

While the correlations uncovered here are plausible,these data must be considered preliminary. The relia-bility of most of the survey subscales was only mod-erate, and internal consistency was low for both theEOLC attitudes and emotions subscales. This impre-cision in the survey instrument may have preventedsignificant correlations from being found. However,even with this limitation, a number of modest corre-lations did emerge. The validity of the subscale forperception of overall educational quality may be lim-ited by the use of a single-item measure. Although thisstudy achieved a high response rate, generalizeabilityis limited by inclusion of students from only a singleinstitution. Replication of the study at other institu-tions would be useful once the survey scales have beenimproved. No information is available about the sur-vey nonresponders, although their numbers are smalland it is unlikely that inclusion of data from themwould significantly alter the results.

These data are based on student reports and therewas no objective determination of the presence, in-tensity, or absence of curricular discordances. Theo-retically, individual students who did not report dis-cordance may have had ideal clinical experiences ornot been attuned to discordances that were actuallypresent. However, the preponderance of evidence anddiscussions with a subset of study subjects suggeststhat discordance was in fact common. Although all stu-dents were exposed to the same formal curriculum inEOLC and it is consistently highly-rated by studentsat UCSF, the formal curriculum was not analyzed al-lowing for the possibility that variations in student’s

RABOW ET AL.764

interpretation of the formal curriculum might have in-fluenced the correlations uncovered.

These data are descriptive and do not permit conclu-sions about the direction of influence or causality. Thepreliminary results of an ongoing qualitative researchproject suggests the interpretations offered here are valid.Students invariably describe curricular discordance asmeaning that informal experiences did not meet the ex-pectations they developed from their classroom learningand that they witnessed clinical behavior not meeting thestandards described in lectures. At UCSF, the materialpresented on EOLC in the first 2 years in the formal cur-riculum is highly regarded by students; lectures in thisarea are regularly the course’s most highly rated and thelecturers are regularly nominated for awards in cate-gories such as “Most Inspirational Lecture.”

Although the qualitative data suggest that seeingcurricular discordance is in fact disturbing to students,it is possible that the direction of influence may travelin the opposite direction or be bidirectional. Studentswho already feel that their school does not value EOLCor teaches it badly may be more likely to pick up ondifferences and inconsistencies between what is saidand what is practiced. Students particularly troubledby the institution’s values may have difficulty identi-fying with their school and profession and remainmore cognizant of the differences between what theywere told was appropriate care and how care is actu-ally practiced. Negative perceptions and experiencesmay keep students from enculturating as completelyas those whose experiences are more positive. Futureresearch is needed to further explore the phenomenonof curricular discordance and its impact. A follow upsurvey of the study participants described here at theend of their final year of school is underway to try toassess further the impact of curricular discordance,changes over time, and its influence on actual clinicalprofessional behavior.

Ultimately, to improve EOLC clinical practice, wemust address the hidden influences, strained relation-ships, and unaddressed conflicts which threaten stu-dents’ ability to realize the highest ideals of profes-sionalism.5,36,37 Clearly, faculty must do better.Important efforts are being made at Indiana Univer-sity School of Medicine and other medical schools tocreate institutional culture that supports the values es-poused in formal curriculum.8 While improving cur-ricular consistency is a central goal, given the com-plexity of medical education and socialization, anotherprofound challenge of professionalism may well beempowering students to deliberate, discern and nego-tiate between conflicting messages about how physi-cians are to act.

ACKNOWLEDGMENTS

Our appreciation to the following individuals fortheir assistance with the conceptualization of this proj-ect and its analysis: Susan Block, M.D., Carrie Chen,M.D., Carol Hodgson, Ph.D., Patricia O’Sullivan,Ph.D., and David Irby, Ph.D. Our deepest thanks toFred Hafferty, Ph.D. for his helpful suggestions, in-spiration, and leadership. The following organizationsprovided instrumental financial and technical support:The George Soros Foundation Project on Death inAmerica, the UCSF Teaching Scholar’s Program andOffice of Medical Education, the UCSF Academy ofMedical Educators, and the UCSF Fellowship in Med-ical Education Research.

REFERENCES

1. Hafferty FW, Franks R: The hidden curriculum, ethicsteaching, and the structure of medical education. Acad Med1994;69:861–871.

2. Hundert EM, Hafferty F, Christakis D: Characteristics ofthe informal curriculum and trainees’ ethical choices. AcadMed 1996;71:624–642.

3. Stern DT: In search of the informal curriculum: When andwhere professional values are taught. Acad Med 1998;73(10 Suppl):S28–30.

4. Feudtner C, Christakis DA, Christakis NA: Do clinicalclerks suffer ethical erosion? Students’ perceptions of theethical environment and personal development. Acad Med1994;69;670–679.

5. Hafferty FW: Beyond curriculum reform: Confronting med-icine’s hidden curriculum. Acad Med 1998;73:403–437.

6. Ludmerer KM: Learner-centered medical education. NEngl J Med 2004;352:1163–1164.

7. Redinbaugh EM, Sullivan AM, Block SD, et al: Doctors’emotional reactions to recent death of a patient: Cross sec-tional study of hospital doctors. BMJ 2003;327:185.

8. Suchman AL, Williamson PR, Litzelman DK, Frankel RM,Mossbarger DL, Inui T: Toward an informal curriculumthat teaches professionalism: Transforming the social en-vironment of a medical school. J Gen Intern Med2004;19:501–504.

9. Solomon MZ, O’Donnell L, Jennings B, Guilfoy V, WolfSM, Nolan K, Jackson R, Koch-Weser D, Donnelley S:Decisions near the end of life: Professional views on life-sustaining treatments. Am J Public Health 1993;83:14–23.

10. Baldwin DC, Daugherty SR, Rowley BD: Unethical andunprofessional conduct observed by residents during theirfirst year of training. Acad Med 1998;73:1195–1200.

11. Stern DT: Practicing what we preach? An analysis of thecurriculum of values in medical education. Am J Med1998;104:569–575.

12. Ginsburg S, Regehr G, Stern D, Lingard L. The anatomyof the professional lapse: bridging the gap between tradi-

CURRICULAR DISCORDANCE IN END-OF-LIFE CARE EDUCATION 765

RABOW ET AL.766

tional frameworks and students’ perceptions. Acad Med2002;77:516–522.

13. Ginsburg S, Regehr G, Stern D, Lingard L: To be and notto be: The paradox of the emerging professional stance.Med Educ 2003;4:350–357.

14. Meier DE, Back AL, Morrison RS: The inner life of phy-sicians and care of the seriously ill. JAMA 2001;286:3007–3014.

15. Novack DH, Suchman Al, Clark W, Epstein RM, NajbergE, Kaplan C: Calibrating the physician: Personal awarenessand effective patient care. JAMA 1997;278:502–509.

16. Shanafelt TD, Bradley KA, Wipf JE, Back AL: Burnoutand self-reported patient care in a an internal medicine res-idency program. Ann Intern Med 2002;136:358–367.

17. Hafferty FW. Into the Valley: Death and the Socializationof Medical Students. New Haven, CT: Yale UniversityPress, 1991.

18. Billings JA, Block S: Palliative care in undergraduate med-ical education. JAMA 1997;278:733–738.

19. Barzansky B, Veloski JJ, Miller R, Jonas HS: Education inend-of-life care during medical school and residency train-ing. Acad Med 1999;74(10 Suppl):S102–104.

20. Sullivan AM, Lakoma MD, Block SD: The status of med-ical education in end-of-life care: A national report. J GenIntern Med 2003;18:685–695.

21. Sullivan AM, Warren AG, Lakoma MD, Liaw KR, HwangD Block SD: End-of-life care in the curriculum: A nationalstudy of medical education deans. Acad Med 2004;79:760–768.

22. Wear D: “Face-to-face with It”: Medical students’ narra-tives about their end-of-life education. Acad Med 2002;77:271–277.

23. Ratanawongsa N, Teherani A, Hauer KE: Third-year med-ical students’ experiences with dying patients during theinternal medicine clerkship: A qualitative study of the in-formal curriculum. Acad Med 2005;80:641–647.

24. Institute of Medicine Committee on Care at the End of Life;Field MJ, Cassell CK (eds): Approaching Death: Improv-ing Care at the End of Life. Washington, D.C.: NationalAcademy Press, 1997.

25. Barnard D, Quill T, Hafferty FW, Arnold R, Plumb J, Bul-ger R, Field M: Preparing the ground: Contributions of thepreclinical years to medical education for care near the endof life. Working Group on the Pre-clinical Years of the Na-tional Consensus Conference on Medical Education forCare Near the End of Life. Acad Med 1999;74:499–505.

26. Weissman DE: End-of-life physician education: Is changepossible? J Palliat Med 1998;1:401–407.

27. Fins JJ, Gentilesco BJ, Carver A, Lister P, Acres CA, PayneR, Storey-Johnson C: Reflective practice and palliative careeducation: A clerkship responds to the informal and hid-den curricula. Acad Med 2003;78:307–312.

28. Lloyd-Williams M, Dogra N: Attitudes of preclinical med-ical students towards caring for chronically ill and dyingpatients: Does palliative care teaching make a difference?Postgrad Med J 2004;80:31–34.

29. Coulehan J, Williams PC: Vanquishing virtue: The impactof medical education. Acad Med 2001;76:598–605.

30. Haidet P, Dains JE, Paterniti DA, Hechtel L, Chang T,Tseng E, Rogers JC: Medical student attitudes toward thedoctor-patient relationship. Med Educ 2002;36:568–574.

31. Kitchen AD: Second-year medical students’ experiencewith death among their families and friends. J Med Educ1986;61:760–761.

32. Embertson S, Reise S: Item Response Theory for Psychol-ogists. Mahwah, NJ: Lawrence Erlbaum Associates Pub-lishers, 2000. 33. Molenaar IW: Preface. In: Fischer GH,Molenaar IW (eds): Rasch Models: Foundations, RecentDevelopments, and Applications. New York: Springer,1995.

34. Orlander JD, Fincke BG, Hermanns D, Johnson GA: Med-ical residents’ firs clearly remembered experiences of giv-ing bad news. J Gen Intern Med 2002;17:825–831.

35. Williams CM, Wilson CC, Olsen CH: Dying, death, andmedical education: Student voices. J Palliat Med 2005;8:372–381.

36. ABIM Foundation. American Board of Internal Medicine;ACP-ASIM Foundation. American College of Physicians-American Society of Internal Medicine; European Federa-tion of Internal Medicine: Medical professionalism in thenew millennium: A physician charter. Ann Intern Med2002;136:243–246.

37. Haidet P, Stein HF: The role of the student-teacher rela-tionship in the formation of physicians. The hidden cur-riculum as process. J Gen Intern Med 2006;21:S16–20.

Address reprint requests to:Michael Rabow, M.D.

UCSF1701 Divisadero Street #500

San Francisco, CA 94143-1732

E-mail: [email protected]

CURRICULAR DISCORDANCE IN END-OF-LIFE CARE EDUCATION 767

APPENDIX A. SURVEY ITEMS FOR EACH STUDY DOMAIN

A. Demographic characteristics (including age, gender, ethnicity, and personal experience with the deathof a loved one)

1. What is your age? _____ years

2. What is your gender Female Male

3. What is your ethnicity? African American Native AmericanAsian/Pacific Islander WhiteHispanic Other

4. In your personal life, have you had one or more people close to you die? Yes No

2. Informal curricular experiences (including interactions with patients, role models, and evaluation &feedback)1. As a medical student at UCSF, I have observed interactions with a dying patient.

Never Once Rarely Sometimes Frequently

2. As a medical student at UCSF, I have had interactions with or helped care for a dying patient.

Never Once Rarely Sometimes Frequently

3. Which of the following activities have you observed an Attending or House staff (intern/resident) do?Which have you done yourself or helped someone do? (Please check ALL that apply for each activity)

Observed Observed Done yourself Attending Housestaff or Helped Do

Breaking bad news to a patient . . . . . . . . . . . . . . . . . . . . . Conducting an advance directive discussion . . . . . . . . . . . . Managing physical symptoms at the end of life . . . . . . . . . Managing patients’ emotional suffering at the end of life . . Pronouncing a patient dead . . . . . . . . . . . . . . . . . . . . . . . .

4. In general, when caring for a dying patient or after a patient had died, how often did your ATTEND-INGS or HOUSE STAFF (interns or residents) talk about the experience?

Attendings:Never Rarely Sometimes Frequently Always

House staff:Never Rarely Sometimes Frequently Always

5. In general, which of the following issues did your ATTENDINGS and HOUSE STAFF talk about?(Please check ALL that apply in each column)

Attendings House staffMedical details about the case Medical details about the caseEthical issues Ethical issuesEmotional issues for the patient or family Emotional issues for the patient or familyEmotional issues for your or other clinicians Emotional issues for you or other cliniciansThe attending’s own emotional issues The house staff’s own emotional issues

6. How often have you received feedback from supervisors on your interactions with dying patients?

Never Rarely Sometimes Most of the Time Always Not Applicable

C. Perception of discordance between the formal and informal curriculumStrongly Strongly

Disagree Disagree Neutral Agree Agree

1. There is discordance between what I wastaught in class about end-of-life care and what I am expected to do on clinical rotations . . 1 2 3 4 5

2. There is discordance between what I was taught in class about end-of-life care and how I have seen my attendings behave . . . . . . . . . . . 1 2 3 4 5

3. There is discordance between what I was taught in class about end-of-life care and how I have seen house staff behave . . . . . . . . . . . . . . . . . . 1 2 3 4 5

4. There are unwritten and unspoken rules that govern how medical students should behavewhen caring for dying patients . . . . . . . . . . . . . 1 2 3 4 5

D. EOLC attitudesStrongly Strongly

Disagree Disagree Neutral Agree Agree

1. All physicians should become competent in providing end-of-life care . . . . . . . . . . . . . . . . . 1 2 3 4 5

2. My development as a person has been enriched by my medical school end-of-life care experiences. . . . . . . . . . . . . . . . . . . . . . . . 1 2 3 4 5

3. If given a choice, I prefer to minimize contact with dying patients during my medical training. . . 1 2 3 4 5

4. It is appropriate for medical students to learn history and physical exam skills on patients nearing the end of life . . . . . . . . . . . . . . . . . . . 1 2 3 4 5

E. EOLC emotions1. Regarding your experiences learning about, observing, or interacting with patients at the end of life,

overall, how easy or difficult were the experiences for you emotionally?

Very Difficult Somewhat Difficult Neutral Somewhat Easy Very Easy

Strongly Strongly

Disagree Disagree Neutral Agree Agree

2. Within my role as a third-year medical student, I feel emotionally prepared to provide care to dying patients . . . . . . . . . . . . . . 1 2 3 4 5

3. At some point during medical school, I have felt badly about how I have seen dying patients treated . . . . . . . . . . . . . . . . . . . . 1 2 3 4 5

4. At some point during medical school, I have had reservations or felt guilty about how I have treated or interacted with patients . . . . . . . 1 2 3 4 5

5. I have felt particularly guilty about how I have treated patients who were nearing the end of life . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 3 4 5

RABOW ET AL.768

CURRICULAR DISCORDANCE IN END-OF-LIFE CARE EDUCATION 769

F. Perception of educational quality1. How would you rate the overall quality of your end-of-life care experiences at UCSF thus far?

Outstanding Good Average Poor Terrible

G. Regard for institutional EOLC valuesStrongly Strongly

Disagree Disagree Neutral Agree Agree

1. At UCSF overall, end-of-life care is highly valued . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 3 4 5

2. UCSF Medical School Faculty are sensitive to the possibility that students might have emotional reactions to caring for dying patients . . 1 2 3 4 5

3. From what I can tell, patients who die at UCSF are cared for in a manner I would wish for myself, my family, or my friends . . . . 1 2 3 4 5

EOLC, end-of-life care.